J. Org. Chem. 2001, 66, 7067-7071
7067
Self-Sensitized Photooxygenation of 3,4-Dialkoxyfurans to Vitamin C or Its Derivatives Gholam H. Hakimelahi,*,† Moti L. Jain,† Tai Wei Ly,† I-Chia Chen,† Krishna S. Ethiraj,† Jih Ru Hwu,† and Ali A. Moshfegh‡ Institute of Chemistry, Academia Sinica, Taipei, Taiwan 115, Republic of China, and Department of Pharmacology, Shiraz University, Shiraz, Iran
[email protected] Received May 24, 2001
Self-sensitized photooxygenation of 3,4-dialkoxyfurans 3a-d with molecular oxygen and UV- or sunlight at room temperature gave vitamin C derivatives 2a-d in good to excellent yields. Furan 3c, having photodegradable protecting groups, was also photooxygenated to give L-ascorbic acid (1) in a “one-pot” reaction. Furthermore, a novel photolytic transformation was developed for deuteration of furan 3b at the C-2 position with D2O to give furan 3d in 95% yield. Toxicity of furans 3a-c and butenolides 2a-c against human embryonic cell, murine embryo fibroblasts, normal fibroblasts, HeLa, and Vero cell lines in the presence of oxygen and indirect solar light was found to be much less than those of the antipsoriasis drugs anthralin and 8-methoxypsoralen. Introduction 3,4-Dialkoxyfurans are known to be much more reactive than furan itself.1 Quantum mechanical calculations by PPP- and CNDO-methods showed that, in comparison to furan, 3,4-dialkoxyfurans possess enhanced partial negative charge at the C-2 and C-5 positions, higher acidity at the R-positions, a stronger and, at the same time, inverted dipole moment, and high reactivity even toward electrophiles.1 As such, it occurred to us that 3,4dialkoxyfurans may also readily react with oxygen to serve as efficient antioxidants. Thus, suitable analogues may be useful as therapeutic agents. Psoriasis is a highly proliferative skin disease;2a its topical treatment often involves anthralin2b or psoralens2c (i.e., 8-methoxypsoralen, 8-MOP). Psoriatic cells have been shown to consume more molecular oxygen than normal epidermal cells,3 and so drugs such as anthralin act as oxygen scavengers within the cell and thus lower the proliferative rate.4 However, the generation of superoxide anion, a highly toxic species,5 by anthralin inside the cell is a source of serious side effects.6 On the other hand, 8-MOP, administered orally or topically, is acti* To whom correspondence should be addressed. Phone: (886)-22789-8682. Fax: (886)-2-2783-1237. † Academia Sinica. ‡ Shiraz University. (1) Iten, P. X.; Hofmann, A. A.; Eugster, C. H. Helv. Chim. Acta 1978, 61, 430. (2) (a) Ramsay, D. L.; Hurley H. J. In Dermatology; Moschella, S. L., Hurley H. J., Eds.; W. B. Saunders: Philadelphia, 1985; Vol. 1, pp 499-556. (b) Ashton, R. E.; Andre, P.; Lowe, N. J.; Whitefield, M. J. Am. Acad. Dermatol. 1983, 9, 173. (c) Acqua, F. D.; Jori, G. Photochemistry. In Principles of Medicinal Chemistry; Foye, W. O., Lemke, T. L., Williams, D. A., Eds.; Williams & Wilkins: London, 1995; pp 893-907 and references cited therein. (d) Roelandts, R. Arch. Dermatol. 1984, 129, 662. (3) Hammer, H.; Hellerstorm, C. Acta Derm. Venereol. (Stockholm) 1968, 48, 563. (4) Bruce, J. M.; Kerr, C. W.; Dodd, N. J. F. J. Chem. Soc., Faraday Trans. 1 1987, 83, 85. (5) Singh, A. CRC Handbook of Free Radicals and Antioxidants in Biomedicine; Miquel, J., Quintanilha, A., Weber, A. Eds.; CRC: Boca Raton, Florida, 1989; Vol 1, pp 17-36. (6) Muller, K.; Gurster, D.; Piwek, S.; Wiegrebe, W. J. Med. Chem. 1993, 36, 4099.
Scheme 1. Preparation of 3,4-Dialkoxyfurans 3a-c from L-Ascorbic Acid (1)
vated by UV light which the patient is subjected to after administration. Photosensitization, and consequently antiproliferative effects, occur when psoralen reaches the skin, either by local absorption following topical application or through the circulation after ingestion and absorption through the gastrointestinal tract.2c However, mutagenic and photocarcinogenic properties related to these drugs are undesired side effects.2c The immune system, the Langerhans cells, and circulating T-lymphocytes may also be affected.2d Herein we report our findings that 3,4-dialkoxyfurans 3a-d can function as efficient oxygen scavengers, yet do not produce superoxide. Newly synthesized furans 3a-c and their photooxygenated products 2a-c exhibited much less cellular toxicity than either anthralin or 8-MOP. Results and Discussion 3,4-Dialkoxyfurans 3a-c were prepared from L-ascorbic acid (vitamin C, 1) in three steps as shown in Scheme 1. After sequential protection of the hydroxy groups in
10.1021/jo010531l CCC: $20.00 © 2001 American Chemical Society Published on Web 09/27/2001
7068
J. Org. Chem., Vol. 66, No. 21, 2001
Scheme 2. Mechanism for the Conversion of Dialkoxyfurans 3 to Butenolides 2
1, the resultant butenolides7 2a-c were treated with diisobutylaluminum hydride in toluene at -78 °C to give the desired furans 3a-c (∼82% yield). Upon irradiation with UV light (λ > 250 nm) under an oxygen atmosphere at room temperature, 3a and 3b in CDCl3, MeOH, or 65:35 (mL/mL) MeOH/H2O were respectively converted to butenolides 2a and 2b in 7076% yields within 3 h (Scheme 2). The optical rotations of 2a ([R]21D 39.869°, c ) 1.17 g/100 mL, CHCl3) and 2b ([R]23D 9.758°, c ) 1.00 g/100 mL, CHCl3) obtained in this way were in good agreement with those synthesized directly from L-ascorbic acid (1) (Scheme 1). Photochemical conversion of triplet molecular oxygen (3O2) to singlet oxygen (1O2) often requires a sensitizer, such as rose bengal.8 Although the conversion of 3a,b to 2a,b occurred in the presence of rose bengal, it was discovered that the photosensitizer was not necessary. This result is interesting since furans acting as sensitizers is not hitherto described. The ability of furans 3a and 3b to act as sensitizers may be due to the presence of electrondonating groups at C-3 and C-4 positions, which facilitate the intersystem crossing of an excited singlet to the triplet and therefore generation of 1O2. To prove this, we irradiated, individually, 3a and 3b in the presence of oxygen and a singlet oxygen quencher such as 1,4diazabicyclo[2.2.2]octane (DABCO). The reactions failed to produce the corresponding butenolides 2a and 2b, and only starting materials were recovered. Moreover, we found that the alkoxy groups in furans were essential for their conversions to butenolides, as evidenced in the unsuccessful photooxygenation of 2,5-dimethylfuran and dimethyl 3,4-furandicarboxylate, along with furfuryl (7) (a) Jung, M. E.; Shaw, T. J. J. Am. Chem. Soc. 1980, 102, 6304. (b) Mei, N.-M. Ph.D. Thesis, National Tsing Hua University, Hsinchu, Taiwan, 2001. (8) (a) Horspool, W.; Armesto, D. Organic Photochemistry: A Comprehensive Treatment; Ellis Horwood: New York, 1992; pp 1012. (b) Turro, N. J. In Molecular Photochemistry; W. A. Benjamin: New York, 1965; pp 132-133.
Hakimelahi et al.
alcohol and the corresponding tert-butyldimethylsilyl ether derivative under the same reaction conditions. Furthermore, the oxidation of furans 3a and 3b to the respective butenolides 2a and 2b did not take place in the dark even when the reaction temperature was raised to 77 °C. The strong absorption of 3,4-dialkoxyfurans 3a and 3b in the 270-300 nm range suggested that the oxidation might be promoted by sunlight.9a-c This was ascertained by exposing an air-saturated solution of 3a or 3b in CCl4 to sunlight for 8 h. In this way, butenolides 2a or 2b were isolated in 82-87% yield. Replacement of air with oxygen gas allowed for a more efficient reaction; quantitative yields were obtained after 6 h. The efficiency of the oxidation process was found to be solvent dependent. The half-life of 3a was 3.5, 12.0, and ∼75 h for its oxidation to 2a in CCl4, CDCl3, and MeOH, respectively, in which the corresponding lifetimes of 1O2 are 700, 300, and 5.0 µs, respectively.9d Thus, we conclude that the rate of conversions of 3a f 2a and 3b f 2b depends on the lifetime of 1O2 in reaction media. We also measured consumption of molecular oxygen by furan 3b under sunlight. A slow and yet steady uptake of oxygen10 by 3b (0.886 g, 3.63 mmol) in CCl4 (15 mL) proceeded in a catalytic hydrogenator under atmospheric pressure for 6 h. Results from 1H NMR studies indicate paralleled rate of oxygen uptake and the conversion of 3b f 2b. The reaction was complete after a 99% mol equiv of 3b uptake of oxygen gas (22.8 mL) was reached. The above photochemical transformation was then conducted in the presence of nitro blue tetrazolium11 to check for the presence of superoxides. Gratifyingly, the UV absorption at 560 nm associated with nitro blue diformazan, the reaction product between nitro blue tetrazolium and superoxide, was not observed. As well, the conversion of furan 3b to butenolide 2b was also found to occur in the presence of superoxide dismutase in 80% yield. Given these two results, we are confident that the superoxide anion is not involved in these reactions. The conversion of 3 f 2 (85% yield) was also found to be achieved in the presence of a stoichiometric amount of a radical scavenger, 2,6-di-tert-butyl-4-methylphenol, which indicates that radical intermediates are not involved in the above conversions. In addition, the toxicity of furans 3a-c and butenolides 2a-c against various cell lines in the presence of sunlight and oxygen was found to be much less than anthralin and 8-MOP. A proposed mechanism for the photooxidation of 3,4dialkoxyfurans 3 to 2 is depicted in Scheme 2. Under photolytic conditions, furans 3 could convert 3O2 to 1O2, which would be consumed through [4 + 2] cycloaddition12a with furan to form endoperoxides 4.12b,13 Structural (9) (a) Urbach, F. Man and Ultraviolet Radiation. In Human Exposure to Ultraviolet Radiation: Risk and Regulations; Passchier, W. F., Bosnjakovic, B. F. M., Eds.; Elsevier: Amsterdam, 1987; pp 3-17. (b) Moan, J.; Dahlbach, A. Ultraviolet Radiation and Skin Cancer: Epidermiological Data from Scandinavia. In Environmental UV Photobiology; Young, A. R., Bjorn, L. O., Moan, J., Nultsch, W., Eds.; Plenum: New York, 1993; pp 255-293. (c) Gies, H. P.; Roy, C. R.; Toomey, S.; MacLennan, R.; Watson, M. Photochem. Photobiol. 1995, 62, 1015. (d) Bellus, D. In Advances in Photochemistry; Pitts, J. N., Jr.; Hammond, G. S., Gollnick, K., Grosjean, D., Eds.; John Wiley & Sons: New York, 1979; Vol. 11, p 105. (10) Hakimelahi, G. H.; Just, G. Helv. Chim. Acta 1982, 65, 1359. (11) Muller, K. Arch. Pharm. (Weinheim) 1988, 321, 385. (12) (a) Adam, W.; Peters, E. M.; Peters, K.; Prein, M.; von Schnering, H. G. J. Am. Chem. Soc. 1995, 117, 6686. (b)Feringa, B. L. Recl. Trav. Chim. Pays-Bas 1987, 106, 469 and references therein. (13) Gollnick, K.; Griesbeck, A. Tetrahedron 1985, 41, 2057.
Self-Sensitized Photooxygenation
J. Org. Chem., Vol. 66, No. 21, 2001 7069
Table 1. Toxicity of Furans 3a-c, Butenolides 2a-c, Rose Bengal, Anthralin, and 8-Methoxypsoralen (8-MOP) in the Presence of Solar Light and Oxygen on the Growth of Cell Lines CC50a (µM)
a
compound
HEL
MEF
Hef522
HeLa
Vero
2a 2b 2c 3a 3b 3c rose bengal anthralin 8-MOP
>120 >120 >120 83.25 ( 3.14 93.04 ( 2.87 87.69 ( 1.90 120 >120 >120 70.11 ( 2.61 77.35 ( 3.04 80.46 ( 2.51 120 >120 >120 90.05 ( 2.79 95.76 ( 1.94 98.14 ( 3.02 0.13 ( 0.04 16.05 ( 1.32 7.08 ( 0.55
>120 >120 >120 63.92 ( 1.35 52.63 ( 0.86 39.78 ( 1.15 120 >120 >120 78.51 ( 1.97 84.12 ( 2.13 95.69 ( 1.88 0.08 ( 0.01 13.49 ( 0.76 5.28 ( 0.30
Cytotoxic concentration required to reduce cell growth by 50%.
features of furan endoperoxides govern formation of different products.12 We believe that the C-3 alkoxy group in 4 would induce a C-O bond cleavage to generate oxonium ion 5,14 which undergoes a 1,2-hydride shift15 to afford β,γ-butenolides 6. The oxygen atom that was lost during the conversion of 5 to 6 was transferred to a suitable acceptor such as dimethyl sulfide to produce dimethyl sulfoxide.13,14,16 Finally, the C-C double bond in 6 shifts to the thermodynamically more stable R,βposition through tautomerizations to give vitamin C derivatives 2, a class of nontoxic compounds (see Table 1).17 To observe the endoperoxide intermediate, photooxygenation of 3b at -75 °C in CDCl3/CFCl3 was carried out. After 4 h, the 1H NMR spectrum, recorded at -70 °C, showed the presence of only the endoperoxide 4b. When the solution was allowed to warm-up to room temperature, the spectrum of endoperoxide 4b collapsed to that of butenolide 2b. It should be noted that the reaction of dimethoxyfuran 3b with other standard oxidants such as bromine or 3-chloroperbenzoic acid produced a mixture of unidentifiable compounds. As such, bimolecular oxygen transfer is a critical element of the mechanism. Furan 3c, possessing photodegradable groups,18 was then synthesized as a precursor to vitamin C (1) in a “one-pot” reaction. Thus, irradiation of 3c in a mixture of CCl4 and MeOH (1:1) under an oxygen atmosphere produced ascorbic acid (1) in 74% yield after 7 h. Analysis of the above reaction after 2.0 h showed that butenolide 2c is an intermediate for the formation of 1, as expected. After 3.0 h, we isolated 2,3-bis(4-methoxybenzyl)-5,6-isopropylidene-L-ascorbic acid (2c, 40% yield) and 2,3bis(4-methoxybenzyl)-L-ascorbic acid (30% yield). This indicates that photodegradation of isopropylidene group is faster than that of the 4-methoxybenzyl functionality. On the other hand, the same reaction under direct exposure to sunlight produced butenolide 2c in 78% yield after 16 h. Indirect exposure of furans 3a-c to sunlight in a mixture of DMSO-d6 and D2O (1:1) under oxygen atmosphere also generated butenolides 2a-c in about 80% yield after 42 h. According to the pathway shown in Scheme 2, the C-2 hydrogen in 3 is transferred to the C-4 position of 2. To (14) For furan endoperoxides as oxygen atom donors, see: Adam, W.; Rodriguez, A. J. Am. Chem. Soc. 1980, 102, 404 and references therein. (15) Sorensen, T. S.; Whitworth, S. M. J. Am. Chem. Soc. 1990, 112, 6647. (16) Wasserman, H. H.; Saito, I.; Vinick, F. J. cited In Matsuura, T.; Saito, I. Photochemistry of Heterocyclic Compounds; Buchardt, O., Ed.; Wiley: NewYork, 1976; p 456. (17) Sato, T.; Niino, Y.; Kakegawa, T.; Matsumoto, H. Japan Patent 01 228 989, 1989; Chem. Abst. 1990, 112, 158838. (18) Hakimelahi, G. H. Unpublished results.
Scheme 3. A plausible Mechanism for Photolytic Deuteration of 3,4-Dialkoxyfuran 3b
obtain evidence in support of this mechanism, we irradiated furan 3d bearing a deuterium at the C-2 position in the presence of 3O2 in CDCl3 with UV light (λ > 250 nm). Butenolide 2d bearing a deuterium at the C-4 position was obtained in 90% yield along with butenolide 2b in 10% yield. In less anhydrous conditions, the yield of the deuterated butenolide 2d decreased to 60%, whereas the yield of 2b increased to 40%. These results lend support to the proposed mechanism involving a tautomerization of 6 to 2 (Scheme 2), rather than a photochemically allowed suprafacial 1,3-hydrogen shift. Furthermore, we developed an efficient and intriguing method for the preparation of deuterated furan 3d. Irradiation of 3b in a mixture of CDCl3 and D2O with UV light (λ > 250 nm) in the absence of oxygen gave 3d in 95% yield. The proposed mechanism shown in Scheme 3 can account for the transformation. Cyclopropene intermediate 819a,19b is generated via diradical 7 by photolysis of 3b. Deuterium exchange of 8 with D2O to generate 10 via oxonium ion 9 could result from the electron-donating capability of the vinylic methoxy group. Finally, photoisomerization of 10 will give deuterated furan 3d. Anticellular Activity. Inhibition of the proliferation of human embryonic cell (HEL), murine embryo fibroblasts (MEF), normal fibroblasts (Hef522), HeLa, and (19) (a) Dean F. M. In Advances in Heterocyclic Chemistry; Katritzky A. R., Ed; Academic Press: New York, 1982; Vol. 31, pp 238-239 and references therein. (b) Rendall, W. A.; Torres, M.; Strausz, O. P. J. Org. Chem. 1985, 50, 3034.
7070
J. Org. Chem., Vol. 66, No. 21, 2001
Vero cells by furans 3a-c, butenolides 2a-c, rose bengal, anthralin, and 8-MOP were carried out20 in the presence of oxygen and solar light. The control cell lines received the same amount of oxygen. Toxicity of the tested compounds is expressed as the cytotoxic concentration required to reduce normal cell growth by 50% (CC50). Results are summarized in Table 1. Anthralin generates superoxide anion, a highly toxic species;5,6 yet furans 3a-c did not form superoxide in the photooxidation process. Thus, furans 3a-c were found to be much less toxic than anthralin. As such, the second oxygen atom lost after ring opening of the endoperoxide (cf. step 5 f 6) did not outweigh the benefit of not forming superoxide. On the other hand, oxygen-dependent photodegradation of 8-MOP will produce furocoumarin derivatives, which can react with membrane components, causing irreversible damage to the cell.2c As a result, 8-MOP exhibited higher toxicity than furans 3a-c. Rose bengal as a sensitizer converted 3O2 to 1O2, which damaged the cells (Table 1). Under solar light, furans 3a-c also converted 3O2 to 1O2; yet consumption of which by the furans produced nontoxic butenolides 2a-c. Conclusions 3,4-Dialkoxy-5-alkylfurans 3a-d were developed as novel oxygen scavengers, yet they did not produce noxious superoxide. In the presence of UV- or sunlight and air, photooxygenation of 3a-d proceeded through an unprecedented pathway at room temperature to give ascorbic acid 1 or its derivatives 2a-d in good to excellent yields. Newly synthesized furans 3a-c were found to be much less toxic than anthralin toward various cell lines. These properties may prove to be of great potential to the development of furans 3a-c as antipsoriasis therapeutics.
Experimental Section General. For anhydrous reactions, glassware was dried overnight in an oven at 120 °C and cooled in a desiccator over anhydrous CaSO4 or silica gel. Reagents were purchased from Fluka Chemical Co. and used as is. Solvents, including dry ether and tetrahydrofuran (THF), were obtained by distillation from the sodium ketyl of benzophenone under nitrogen. Other solvents, including chloroform, dichloromethane, ethyl acetate, and hexanes were distilled over CaH2 under nitrogen. Absolute methanol and ethanol were purchased from Merck and used as received. Melting points were obtained with a Bu¨chi 510 melting point apparatus. Infrared (IR) spectra were recorded on a PerkinElmer Paragon 1000 Fourier Transform spectrophotometer. Carbon-13 NMR and proton NMR spectra were obtained on a Varian XL-300 (300 MHz) spectrometer. Mass spectra were carried out on a VG 70-250 S mass spectrometer. Photolytic experiments were carried out by use of a 450-watt mediumpressure mercury Canrad-Hanovia lamp with a Pyrex filter. A sunlight lamp was obtained from Surprise. Com., Inc. Purification on silica gel refers to gravity column chromatography on Merck Silica Gel 60 (particle size 230-400 mesh). Analytical TLC was performed on precoated plates purchased from Merck (Silica Gel 60 F254). Compounds were visualized by use of UV light, I2 vapor, or 2.5% phosphomolybdic acid in ethanol with heating. (20) (a) Rai’c-Mali’c, S.; Hergold-Brundi’c, A.; Nagl, A.; Grdisa, M.; Paveli’c, K.; DeClercq, E.; Mintas, M. J. Med. Chem. 1999, 42, 26732678. (b) Rai’c-Mali’c, S.; Svedruzi’c, D.; Gazivoda, T.; Marunovi’c, A.; Hergold-Brundi’c, A.; Nagl, A.; Balzarini, J.; DeClercq, E.; Mintas, M. J. Med. Chem. 2000, 43, 4806-4811.
Hakimelahi et al. Standard Procedure for Preparation of 5-Substituted3,4-dialkoxyfurans 3a-c. To a solution of butenolides 2a-c (1.0 mmol) in toluene (17.0 mL) was added diisobutylaluminum hydride at -78 °C under argon. After stirring for 1.0 h at the same temperature, phosphate buffer (1.0 M, pH 7.2, 20 mL) solution was added. The organic layer was washed with water (20 mL), dried (MgSO4), filtered, and concentrated under reduced pressure. The residue was purified by use of column chromatography (CH2Cl2 as eluant) to give the corresponding furans 3a-c in about 82% yield. For 3a: 1H NMR (CDCl3) δ 3.73 (dd, J ) 8.00, 6.02 Hz, 1H, HC(7)), 3.84 (dd, J ) 8.00, 6.02 Hz, 1H, HC(7)), 4.79 (dd, J ) 6.02, 6.02 Hz, 1H, H-C(6)), 4.88 (s, 2H, OCH2O), 4.95 (d, J ) 15.60 Hz, 2H, H2COC(4)), 5.04 (s, 2H, H2COC(3)), 6.92 (s, 1H, H-C(2)), 7.30-7.37 (m, 10H, 2C6H5); 13C NMR (CDCl3) δ 66.38, 68.33, 72.98, 75.09, 95.39, 124.10, 127.36, 128.15, 128.25, 128.34, 128.54, 136.30, 136.70, 137.80, 142.99; UV (EtOH): λmax 269 ( 19030), 275 ( 18500), 298 ( 17300); CIMS m/z: 353 (M+ + 1), 352 (M+). Anal. Calcd for C21H20O5: C, 71.58; H, 5.72. Found: C, 71.55; H, 5.74. For 3b: 1H NMR (CDCl3) δ 1.40 (s, 3H, CCH3), 1.46 (s, 3H, CCH3), 3.69 (s, 3H, OCH3), 3.83 (s, 3H, OCH3), 4.04 (dd, J ) 7.00, 5.50 Hz, 1H, HC(7)), 4.11 (dd, J ) 7.00, 5.50 Hz, 1H, HC(7)), 5.07 (dd, J ) 5.50, 5.50 Hz, 1H, H-C(6)), 6.91 (s, 1H, H-C(2)); 13C NMR (CDCl3) δ 25.90, 26.13, 58.11, 61.21, 66.13, 68.94, 109.43, 122.74, 136.31, 138.24, 144.28; UV (EtOH): λmax 267 ( 18330), 274 ( 17410), 296 ( 18400); CIMS m/z: 229 (M+ + 1), 228 (M+). Anal. Calcd for C11H16O5: C, 57.89; H, 7.07. Found: C, 57.91; H, 7.12. For 3c: 1H NMR (CDCl3) δ 1.28 (s, 3H, CCH3), 1.35 (s, 3H, CCH3), 3.68 (s, 3H, OCH3), 3.70 (s, 3H, OCH3), 3.72 (dd, J ) 7.00, 5.45 Hz, 1H, HC(7)), 3.83 (dd, J ) 7.00, 5.45 Hz, 1H, HC(7)), 4.78 (dd, J ) 5.45, 5.45 Hz, 1H, H-C(6)), 4.73 (s, 2H, OCH2), 4.89 (s, 2H, OCH2), 6.86 (s, 1H, H-C(2)), 6.96 (AB, 4H, C6H4), 7.02 (AB, 4H, C6H4); 13C NMR (CDCl ) δ 25.83, 26.10, 55.12, 66.08, 68.72, 72.70, 3 74.70, 109.23, 113.64, 113.85, 124.09, 128.47, 128.95, 129.96, 130.22, 136.59, 137.39, 142.96, 159.55; UV (EtOH): λmax 270 ( 21000), 278 ( 19490), 297 ( 18550); CIMS m/z: 441 (M+ + 1), 440 (M+). Anal. Calcd for C25H28O7: C, 68.17; H, 6.41. Found: C, 68.20; H, 6.51. 2-Deuterio-3,4-dimethoxy-6,7-isopropyliden-5-ylfuran (3d). A solution of 3b (114 mg, 0.491 mmol) in a mixture of CDCl3 (1.6 mL) and D2O (1.6 mL) was irradiated with UV light by use of a medium-pressure mercury lamp (450 W, λ > 250 nm) equipped with a Pyrex glass filter at room temperature under argon for 5 h. The solution contained >95% of 3d as evidenced by NMR analyses. 1H NMR (CDCl3) δ 1.44 (s, 3H, CCH3), 1.49 (s, 3H, CCH3), 3.72 (s, 3H, OCH3), 3.87 (s, 3H, OCH3), 4.10 (dd, J ) 6.96, 6.57 Hz, 1H, HC(7)), 4.17 (dd, J ) 6.96, 6.57 Hz, 1H, HC(7)), 5.14 (dd, J ) 6.57, 6.57 Hz, 1H, H-C(6)); 13C NMR (CDCl3) δ 25.73, 25.92, 58.14, 61.22, 66.16, 68.99, 109.45, 122.54 (t, J ) 125.00 Hz, D-C(2)), 136.29, 138.29, 144.19; UV (EtOH): λmax 266 ( 18,000), 275 ( 16,500), 295 ( 17,600); CIMS m/z: 230 (M+ + 1), 229 (M+). Standard Procedure for Preparation of 2,3-Dialkoxy5,6-alkylidene-L-ascorbic Acids 2a-d. A solution of 3a-d (0.50 mmol) in CDCl3 (2.70 mL) was irradiated by a 450-W medium-pressure Hanovia mercury lamp (λ > 250 nm) placed inside a water cooled (15-25 °C) Pyrex immersion well. Oxygen flow was maintained during photolysis, and the reaction temperature was kept near 20 °C by regulating the temperature of the water. After 3.0 h, the solvent was concentrated under reduced pressure, and the residue was purified by use of column chromatography (20% EtOAc in hexanes as eluant) to give the corresponding vitamin C derivatives 2a-d in about 75% yield. For 2a: mp 62-63 °C; IR (KBr) 1764 (CdO), 1681 (CdC) cm-1; H NMR (CDCl3) δ 3.82 (dd, J ) 8.60, 6.60 Hz, 1H, HC(6)), 3.91 (dd, J ) 8.60, 6.60, 1H, HC(6)), 4.09-4.14 (ddd, J ) 6.60, 6.60, 2.80 Hz, 1H, H-C(5)), 4.46 (d, J ) 2.80 Hz, 1H, H-C(4)), 4.81 (s, 2H, OCH2O), 5.02 (s, 2H, H2COC(2)), 5.09 (d, J ) 8.60 Hz, 2H, H2COC(3)), 7.09-7.28 (m, 10H, 2C6H5); 13C NMR (CDCl3) δ 65.56, 72.98, 73.45, 73.59, 74.64, 96.06, 121.04, 127.62, 128.45, 128.90, 135.13, 135.75, 156.39, 168.66; MS m/z: 368 (M+); CIMS m/z: 369 (M+ + 1), 368 (M+). Anal. Calcd for C21H20O6: C, 68.47; H, 5.47. Found: C, 68.58; H, 5.59. For 2b: mp 91-92 °C; IR
Self-Sensitized Photooxygenation (KBr) 1760 (CdO), 1677 (CdC) cm-1; 1H NMR (CDCl3) δ 1.33 (s, 3H, CCH3), 1.36 (s, 3H, CCH3), 3.82 (s, 3H, H3COC(2)), 3.94 (dd, J ) 8.40, 6.40 Hz, 1H, HC(6)), 4.02 ((dd, J ) 8.40, 6.40 Hz, 1H, HC(6)), 4.12 (s, 3H, H3COC(3)), 4.21-4.29 (ddd, J ) 6.40, 6.40, 3.00 Hz, 1H, H-C(5)), 4.48 (d, J ) 3.00 Hz, 1H, H-C(4)); 13C NMR (CDCl3) δ 25.52, 25.78, 59.43, 60.30, 65.20, 73.83, 74.43, 110.34, 123.20, 156.66, 168.88. MS m/z: 244 (M+), 229 (M+ - CH3), 187 (M+ - CH3 - C(CH3)2); CIMS m/z: 245 (M+ + 1), 244 (M+), 229 (M+ - CH3), 187 (M+ - CH3 C(CH3)2). Anal. Calcd for C11H16O6: C, 54.09; H, 6.60. Found: C, 54.18; H, 6.71. For 2c: Foam; IR (CHCl3) 1763 (CdO), 1679 (CdC) cm-1; 1H NMR (CDCl3) δ 1.33 (s, 3H, CCH3), 1.37 (s, 3H, CCH3), 3.77 (s, 3H, OCH3), 3.78 (s, 3H, OCH3), 3.94 (dd, J ) 8.42, 6.41 Hz, 1H, HC(6)), 4.01 (dd, J ) 8.42, 6.41 Hz, 1H, HC(6)), 4.14-4.20 (ddd, J ) 6.41, 6.41, 2.90 Hz, 1H, H-C(5)), 4.46 (d, J ) 2.90 Hz, 1H, H-C(4)), 4.98-5.10 (m, 4H, 2OCH2), 7.08 (AB, 4H, C6H4), 7.22 (AB, 4H, C6H4); 13C NMR (CDCl3) δ 25.51, 25.75, 54.92, 55.07, 65.09, 73.23, 73.84, 74.54, 109.98, 113.55, 113.78, 120.76, 127.35, 128.05, 129.64, 130.62, 130.74, 131.04, 156.72, 179.78, 169.08; CIMS m/z: 457 (M+ + 1), 456 (M+). Anal. Calcd for C25H28O8: C, 65.78; H, 6.18. Found: C, 65.81; H, 6.22. For 2d: mp 98-102 °C; IR (KBr) 1770 (CdO), 1675 (CdC) cm-1; 1H NMR (CDCl3) δ 1.36 (s, 3H, CCH3), 1.39 (s, 3H, CCH3), 3.84 (s, 3H, H3COC(2)), 4.03 (dd, J ) 9.6, 6.40, 1H, HC(6)), 4.12 (dd, J ) 9.6, 6.40, 1H, HC(6)), 4.09 (s, 3H, H3COC(3)), 4.29 (dd, J ) 6.40, 6.40 Hz, 1H, H-C(5)); CIMS m/z: 246 (M+ + 1), 245 (M+), 230 (M+ - CH3), 188 (M+ - CH3 - C(CH3)2). Observation of Endoperoxide 4b through Photooxygenation of Furan 3b at Low Temperature. A solution of 3b (0.20 mmol) in CDCl3/CFCl3 (1.2 mL, 3:1) was irradiated with a 450-W medium-pressure Hanovia mercury lamp (λ > 250 nm). During the irradiation, oxygen was bubbled through the solution which was kept at -75 °C. After 4.0 h the reaction was complete, and the 1H NMR spectrum recorded at -70 °C showed only the endoperoxide 4b. 1H NMR (CDCl3/CFCl3, -70 °C) δ 1.50 (br s, 6H, 2 CCH3), 3.59 (br s, 6H, 2 OCH3), 4.21 (dd, J ) 9.51, 6.70 Hz, 1H, HC(7)), 4.35 (dd, J ) 9.51, 6.70 Hz, 1H, HC(7)), 5.26 (dd, J ) 6.70, 6.70 Hz, 1H, H-C(6)), 6.24 (s, 1H, H-C(2)). At -27 °C, the 1H NMR spectrum showed the presence of both endoperoxide intermediate 4b and butenolide 2b (∼1:2). At ambient temperature, product 2b was present exclusively.
J. Org. Chem., Vol. 66, No. 21, 2001 7071 Conversions of Furans 3a-c to Respective Butenolides 2a-c in the Presence of Sunlight. To a solution of 3a, 3b, or 3c (0.10 mmol) in CCl4 containing CH3SCH3 (0.065 mmol) was bubbled air for 1.0 min. The 1H NMR spectrum at 25 °C was taken immediately; then the solution was exposed to sunlight. After 8 h, the spectrum of 3a-c changed to that of 2a-c, respectively, and CH3SCH3 was converted to CH3S(O)CH3. In the large scale reactions (2.0 mmol of 3a-c), 2a (82% yield), 2b (87% yield), and 2c (85% yield) were isolated. Photolysis of Furan 3c to Ascorbic Acid (1). A solution of 3c (440 mg, 0.998 mmol) in a mixture of CCl4 (4.0 mL) and MeOH (4.0 mL) was irradiated with UV light by use of a medium-pressure mercury lamp (450 W, λ > 250 nm) equipped with a Pyrex glass filter at room temperature under oxygen for 7 h. Then, the solution was concentrated, and diethyl ether was added to afford a solid. Filtration and crystallization of the solid with EtOH gave ascorbic acid (1) in 74% yield. It was identical with an authentic sample. Filtrate was evaporated, and the resultant 4-methoxybenzaldehyde was also characterized. Toxicity Test Procedure in Vitro. Human embryonic cell (HEL), murine embryo fibroblasts (MEF), normal fibroblasts (Hef522), HeLa, and Vero cell lines were cultured in DMEM supplemented with 10% FBS, 2.0 mM glutamine, 100 U/mL penicillin, and 100 µg/mL streptomycin in a humidified atmosphere with 5% CO2 and 20% O2 at 37 °C. After 24 h, furans 3a-c, butenolides 2a-c, rose bengal, anthralin, and 8-MOP, at various concentrations, were added, and the cultures were exposed indirectly to solar light during the day and to the sunlight lamp at night. The cell numbers of the control cultures as well as those cultures supplemented with the test compounds were determined after 12, 24, 48, 72, and 96 h of growth. The CC50 values were estimated from doseresponse curves compiled from two independent experiments and represent cytotoxic concentration (µM) required to reduce normal cell growth by 50% after 96 h incubation (Table 1).
Acknowledgment. For financial support, we thank Academia Sinica, Shiraz University, and the National Science Council of Republic of China. JO010531L